An optimization method for the automatic adaptation algorithm of the leaf fan

Abstract

The invention discloses an optimization method of an automatic adaptive algorithm. The method comprises the following steps of a, arranging a design variable space, and arranging a solver; b, operating a sample algorithm to call the solver to perform sampling; c, building a sample database, and performing DoE analysis; d, constructing Kriging model, analyzing a solution space, selecting an appropriate optimization algorithm combination, and adjusting internal control parameters of an optimization algorithm to obtain an optimal point predicted in the current Kriging model; e, calling the solver, and evaluating the predicted point; f, judging whether terminal conditions are met or not, and terminating the process and returning an optimization solution if the terminal conditions are met; if not, executing the steps c, d, e and f in sequence. The method has the advantages that a reasonable and more reasonable design scheme is found and the optimization efficiency is improved through low solving frequency; an optimization strategy is automatically selected, Lagrange constraint handling can be performed on multi-target problems, specialist parameter adaptiveness of the internal optimization method can be achieved, the method is easy and convenient to use, and the requirement of a user for mastering the optimization theory is lowered; automatic programming processing optimizing is achieved, and artificial workloads are reduced.

Description

Technical field [0001] The invention relates to the field of computer technology, in particular to an optimization method of an automatic adaptation algorithm. Background technique [0002] The fundamental goal of optimization is to improve R&D capabilities, reduce energy consumption and product operation and maintenance costs, improve user experience, and accelerate market share. The combination of optimization methods and design has become an inevitable trend. [0003] The optimization of traditional design is through manual trial algorithm, according to design requirements, referring to the design experience of similar products, combining certain theoretical judgments to select design parameters, and then verifying through computer-aided engineering technology, etc., repeat the above process until obtained More satisfactory design so far. However, the inefficiency of manual processing methods and the lack of product quality in this traditional design optimization have promoted...

AI Technical Summary

Problems solved by technology

[0005] Existing optimization methods have their own limitations. Each optimization method has its own scope of application, parameter setting, and internal strategy implementation scheme. If it relies on a certain optimization method alone, or simply collects existing optimization methods , and then the user selects the appropriate method according to the needs. This manual process leads to the user's own optimization experience and professional knowledge as an important factor in the selection of the optimization method. Through this subjective selection, it is impossible to determine whether the selected one is correct or not. As a better method, the quality and efficiency of optimization cannot be determined

[0006] To sum up, traditional optimization methods have high requirements for users, and users need to have certain knowledge of optimization methods, such as selection of methods, parameters, and conditions for using agents. Continuous improvement, such as robustness and adaptability, but all rely on a certain optimization method, or directly select multiple optimization methods to design at the same time, cannot obtain higher efficiency, resulting in high design cost and large amount of calculation; The derivation process is complicated, and a better solution or result cannot be obtained stably

Images